Journal of the Virtual Explorer

Crystallization of amphibole and dolomite in the HP Ulten garnet peridotites due to influx of aqueous-carbonic fluids makes these rocks a relevant natural laboratory to understand C and volatile recycling in the mantle via subduction. This represents the "missing link" of global volatile cycles due to inaccessibility of subduction environments. Also, unravelling "how" C-O-H compounds occur in the mantle, i.e. their speciation, is necessary to predict the evolution of carbon species during long-term geochemical cycles. Natural mineral assemblages and laboratory experiments demonstrate that the composition of C-O-H fluids in the mantle closely relates to the oxidation state of the system (Van Roermund et al., 2002; Carswell and Van Roermund, 2005; Scambelluri et al., 2008; Tumiati et al., 2009; Malaspina et al., 2009b; Malaspina et al., 2010). The latter issue is relevant for C recycling, as either diamond/graphite generally crystallize depending on the chemical potential of oxygen and therefore on the oxygen fugacity. Graphite and diamond generally behave as refractory phases, whereas carbonates, although refractory, are highly reactive (Canil, 1990). An estimate of oxygen fugacity (ƒO₂) attained by the Ulten Zone garnet (+ amphibole + dolomite)-bearing peridotites is thus essential to model C transfer during subduction.

The equilibria between ferric (Fe³⁺) and ferrous (Fe²⁺) iron in mineral assemblages buffer the oxygen fugacity and the C-O-H fluid speciation. On the other hand, the C-O-H species in fluids or melts can control the oxidation state of mantle phase assemblages by redox processes (Luth, 1999). Being the Fe³⁺/Fe²⁺ equilibria of mantle parageneses and the speciation of C-O-H fluids so related, Malaspina et al. (2009b) determined the iron oxidation state of garnet in the Ulten Zone peridotites. Several redox equilibria involve Fe³⁺-garnet components; for an olivine-orthopyroxene-Fe³⁺-garnet assemblage, one of these reactions is:

(1) 2 Fe²⁺₃Fe³⁺₂Si₃O₁₂ (skiagite) = 4 Fe²⁺₂SiO₄ (fayalite) + 2 Fe²⁺₂Si₂O₆ (ferrosilite) + O₂ where the ferric garnet component "skiagite" contains both Fe³⁺ and Fe²⁺ (Luth et al., 1990; Gudmundsson and Wood, 1995; Woodland and Peltonen, 1999). The electron microprobe analysis of the Fe³⁺ content of garnet from the Ulten peridotite (equivalent to skiagite substitution for almandine) adopting the "flank method" (Höfer and Brey, 2007; see details in Malaspina et al., 2009b) provided skiagite contents in the order of 3 (±1) mol%. The corresponding ƒO₂ has been calculated by Malaspina et al. (2009b) using the Perple_X software (Connolly, 1990) applying the redox equilibrium (1) and integrating solid solution models for garnet (pyrope-grossular-almandine-skiagite; Malaspina et al., 2009b), olivine (forsterite-fayalite; Holland and Powell, 1998) and orthopyroxene (enstatite-ferrosilite; Holland and Powell, 1996). The resulting redox equilibria are shown in the isobaric T-log ƒO₂ section computed at 3 GPa pressure (Fig. 14): the dashed curves represent the equilibria selected for pseudocompounds approaching the investigated mineral compositions (olivine = Fo89; orthopyroxene = En89; garnet = 70% pyrope, 10% grossular and variable skiagite/almandine depending on oxygen fugacity). The equilibria involving hematite, magnetite, fayalite, quartz, ferrosilite, wustite and iron at various buffering conditions are also shown in Figure 14 (continuous lines) for comparison. The estimated oxygen fugacities range from -10.69 to -8.94 log units (Fig. 14); the resulting Δlog ƒO₂ [=log ƒO₂ (sample) - FMQ] range between zero and +2 (Fig.15).

Figure 14. Log ƒO₂ and speciation of a graphite–C–O–H fluid.

[View full size]

Log ƒO₂ (a) and speciation of a graphite–C–O–H fluid (b) calculated at 3 GPa and 850 °C as a function of X(O) = n_O/(n_O+n_H). The green line (a) and dashed green area (b) is the ƒO₂ range calculated for the Ulten peridotites. Calculation have been performed with the programme "species" in Perplex package (Connolly, 1995). FMQ = fayalite–magnetite–quartz buffer, WM = wustite–magnetite buffer, IW = iron–wustite buffer.

In Figure 15, the estimated oxygen fugacities of the Ulten Zone garnet peridotites (expressed as Δlog ƒO₂ values) are plotted as a function of P, together with other mantle peridotites from distinct geological environments. Measured ƒO₂ values vary from: (i) 1 to 3 log units below the FMQ in abyssal peridotites (Bryndzia & Wood, 1990); (ii) Δlog ƒO₂ = +1 to –2 for peridotite massifs, with samples from Beni Bousera reaching up to Δlog ƒO₂= –4 (Woodland et al., 1992; Woodland et al., 2006); (iii) Δlog ƒO₂= –1 to +2 for spinel peridotites from subcontinental lithospheric mantle, with significant heterogeneities (Wood et al., 1990; Canil et al., 1990; Brandon and Draper, 1996; Parkinson and Arculus, 1999). The garnet peridotites from sub-cratonic mantle have the lowest ƒO₂ values, below Δlog ƒO₂= –2 (Woodland and Koch, 2003). Differently from these mantle rocks, the Ulten peridotites display high Δlog ƒO₂ values (0 to +2). These values are comparable with those estimated for spinel peridotite xenoliths from subduction settings (Simcoe, Vanatu and Japan; Δlog ƒO₂ between –1 and +1.5; Brandon & Draper, 1996; Parkinson et al., 2003). Importantly similarly high values of ƒO₂ are shown by other UHP orogenic garnet peridotites, like the ones from Sulu (China) (Malaspina et al., 2009b; Malaspina et al., 2010). These data point to a lateral oxygen fugacity variation related to different tectonic settings (e.g. Woodland and Koch, 2003; Frost and McCammon, 2008).

As previously mentioned, the ƒO₂ recorded by the Ulten peridotites have implications on the speciation of the coexisting C-O-H fluid phase. According to Holloway and Reese (1974) and to the model of Connolly (1995), C-saturated fluids can be either methane-bearing or carbon-dioxide bearing as a function of the oxygen fugacity imposed by the redox conditions of the system. At high pressures H₂O is the dominant species in the C–O–H fluid, which will be a mixture of CH₄–H₂O at lower ƒO₂ and of H₂O–CO₂ at higher ƒO₂. Despite the absence of diamond and graphite in the studied peridotites, a C-saturated system can be used as a proxy to estimate the fluid composition. This is possible because at the high pressure and temperature conditions attained by the Ulten Zone garnet-peridotites, the graphite saturation boundary approaches the binary joins H₂O–CO₂ and H₂O–CH₄ (Holloway & Reese, 1974). Based on this reasoning and on calculations performed using the GC-O-H-MRK equation of state by Connolly and Cesare (1993) for a C-saturated C-O-H fluid, Malaspina et al., (2009) have evaluated the C–O–H molecular species concentrations of the metasomatic fluid. This points to high water content of the fluid and the presence of CO₂ (instead of CH₄) at Δlog ƒO₂ > 0. This confirms the expectations of what discussed by Rampone and Morten (2001) and Sapienza et al. (2009) on the basis of petrographic evidence and trace element budgets.

Figure 15. ΔFMQ versus pressure.

Range and average value of oxygen fugacities relative to FMQ (ΔlogƒO₂ = logƒO₂sample - logƒO₂FMQ) for the Ulten garnet peridotites (red line) plotted as a function of pressure (modified after Malaspina et al., 2009b). This is compared with selected examples of spinel and garnet peridotites from different tectonic settings, equilibrated at similar T conditions (~800-900 °C): oceanic mantle lithosphere (Bryndzia & Wood, 1990), peridotite massifs (Woodland et al., 1992; Woodland et al., 2006), garnet peridotite xenoliths from sub-cratonic mantle (Woodland and Koch, 2003) and mantle wedge from spinel facies (Brandon and Draper, 1996; Parkinson and Arculus, 1999; Peslier et al., 2002). Mantle wedge-derived orogenic garnet peridotites from Sulu (Eastern China) and Bardane (Western Gneiss Region) are also reported in black lines (Malaspina et al., 2009b, 2010). The CCO oxygen buffer (C+O₂ = CO₂) and the C-H₂O join (X(O) = 1/3) separating CH₄- and CO₂-rich aqueous fluid, calculated at 900 °C, are also portrayed in dashed lines.